Quality of care is often best assessed from patient outcome. To properly evaluate patient outcome, it is necessary to consider the prognosis of the patient prior to treatment i.e. the patient's severity of illness. Usually severity of illness measures are derived from the opinion of clinical experts. The objective of the proposed study is to derive severity measures from clinical data bases for patients undergoing coronary artery revascularization procedures (coronary artery angioplasty and coronary artery bypass surgery). The severity of illness measure used will be the probability of an adverse outcome as estimated by a logistic regression equation. The value of this model for measuring severity of illness will be assessed by comparing the equation-based severity measures to other severity measures and testing the value of the equation-based severity measures to a Peer Review Organization. There are three phases to this study. In the first phase the applicants will complete their development of equation-based severity measures using detailed clinical data that were collected by the Health Care Financing Administration (HCFA) on 33303 Medicare revascularization patients and are currently resident on the applicant's computer. In the second phase data on 1800 Medicare revascularization patients from a local private hospital specializing in heart disease will be utilized for three purposes: 1) to test the equation-based severity measures derived in the first part of the study. 2) to compare these severity measures with measures derived from the opinion of clinical experts, and 3)_ to improve these severity measures using clinical information that was not available in the HCFA study. In the third phase of the study the applicants will analyze data on 3200 Medicare revascularization patients to assess the value to the Wisconsin Peer Review Organization of the equation-based severity measures developed in phase two of the study. If the equation-based severity measures are shown to be of value in this study, they may also be of value in assessing the quality of care for other types of patients.